Note:
If you use SpaceFold in published research, please cite:
Niec, R. E., Chu, T., Schernthanner, M., Gur-Cohen, S. et al. Lymphatics act as a signaling hub to regulate intestinal stem cell activity. Cell Stem Cell 2022.
https://doi.org/10.1016/j.stem.2022.05.007
Introduction
- SpaceFold maps cell type fraction and cell type specific gene expression profile deconvovled by BayesPrism onto a one-dimensional axis that recapitulates the underlying tissue structure.
- SpaceFold assumes that 1) the tissue has an underlying 1D structure; 2) it is a stereotypical structure that is repeatedly sampled by ST; 3) the cell type fraction of a spatial spot is sufficient in inferring its physical coordinate along the 1D axis.
In this tutorial we will demonstrate how to use SpaceFold to generate gene expression cartography for mouse small intestine tissue. Starting from the BayesPrism output.
Load the BayesPrism package
suppressWarnings(library(SpaceFold))
#> Loading required package: BayesPrism
#> Loading required package: snowfall
#> Loading required package: snow
#> Loading required package: NMF
#> Loading required package: pkgmaker
#> Loading required package: registry
#> Loading required package: rngtools
#> Loading required package: cluster
#> NMF - BioConductor layer [OK] | Shared memory capabilities [NO: bigmemory] | Cores 71/72
#> To enable shared memory capabilities, try: install.extras('
#> NMF
#> ')Load the dataset
The rdata file can be downloaded from
"https://figshare.com/ndownloader/files/36113753"
load("SI.sig.bp.rdata")
ls()
#> [1] "SI.sig.bp"SI.sig.bp
#> Input prism info:
#> Cell states in each cell type:
#> $Paneth
#> [1] "Paneth"
#>
#> $TA
#> [1] "TA"
#>
#> $`bottom zone enterocyte`
#> [1] "bottom zone enterocyte"
#>
#> $`mid zone enterocyte`
#> [1] "mid zone enterocyte"
#>
#> $`Lgr5+ progenitor`
#> [1] "Lgr5+ progenitor"
#>
#> $`goblet cycling`
#> [1] "goblet cycling"
#>
#> $plasma
#> [1] "plasma"
#>
#> $`goblet 1`
#> [1] "goblet 1"
#>
#> $`Lgr5+ stem`
#> [1] "Lgr5+ stem"
#>
#> $`secretory progenitor`
#> [1] "secretory progenitor"
#>
#> $`goblet 2`
#> [1] "goblet 2"
#>
#> $macrophage
#> [1] "macrophage"
#>
#> $`top zone enterocyte`
#> [1] "top zone enterocyte"
#>
#> $tuft
#> [1] "tuft"
#>
#> $`neurons/enteroendocrine`
#> [1] "neurons/enteroendocrine"
#>
#> $`cycling/GC B cell`
#> [1] "cycling/GC B cell"
#>
#> $myofibroblast
#> [1] "myofibroblast"
#>
#> $`stromal 3`
#> [1] "stromal 3"
#>
#> $`cDC/monocyte`
#> [1] "cDC/monocyte"
#>
#> $`blood endothelium`
#> [1] "blood endothelium"
#>
#> $`stromal 1/2`
#> [1] "stromal 1/2"
#>
#> $lymphatic
#> [1] "lymphatic"
#>
#> $`T cell`
#> [1] "T cell"
#>
#> $pDC
#> [1] "pDC"
#>
#> $glial
#> [1] "glial"
#>
#> $`resting B cell`
#> [1] "resting B cell"
#>
#>
#> Identifier of the malignant cell type: NA
#> Number of cell states: 26
#> Number of cell types: 26
#> Number of mixtures: 4487
#> Number of genes: 3908
#>
#> Initial cell type fractions:
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> Min. 0.000 0.000 0.000 0.004 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
#> 1st Qu. 0.005 0.002 0.004 0.115 0.002 0.006 0.002 0.003 0.002 0.006 0.002 0.003 0.029 0.004 0.001 0.001 0.019 0.003 0.005 0.003 0.005 0.002 0.002 0.001 0.001 0.001
#> Median 0.012 0.006 0.010 0.255 0.005 0.014 0.005 0.006 0.004 0.015 0.006 0.004 0.092 0.008 0.002 0.002 0.044 0.008 0.009 0.005 0.011 0.004 0.004 0.002 0.002 0.002
#> Mean 0.050 0.038 0.025 0.282 0.018 0.035 0.013 0.009 0.014 0.036 0.022 0.007 0.196 0.014 0.006 0.003 0.156 0.018 0.013 0.009 0.016 0.007 0.006 0.003 0.003 0.002
#> 3rd Qu. 0.039 0.019 0.025 0.411 0.016 0.045 0.012 0.011 0.013 0.039 0.031 0.008 0.363 0.018 0.005 0.004 0.134 0.022 0.017 0.011 0.022 0.009 0.006 0.004 0.004 0.003
#> Max. 0.778 0.741 0.357 0.859 0.313 0.309 0.352 0.124 0.261 0.349 0.259 0.217 0.794 0.123 0.222 0.066 0.886 0.324 0.127 0.245 0.135 0.171 0.395 0.033 0.109 0.016
#> Updated cell type fractions:
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> Min. 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
#> 1st Qu. 0.004 0.003 0.006 0.063 0.003 0.005 0.002 0.004 0.003 0.005 0.003 0.002 0.035 0.004 0.001 0.002 0.017 0.002 0.003 0.003 0.005 0.002 0.002 0.002 0.001 0.002
#> Median 0.011 0.010 0.016 0.196 0.007 0.011 0.005 0.007 0.007 0.015 0.008 0.004 0.094 0.008 0.002 0.004 0.039 0.006 0.007 0.006 0.011 0.005 0.003 0.003 0.002 0.003
#> Mean 0.049 0.052 0.048 0.227 0.021 0.031 0.013 0.012 0.020 0.039 0.022 0.007 0.199 0.014 0.006 0.005 0.153 0.018 0.011 0.009 0.017 0.008 0.005 0.005 0.005 0.005
#> 3rd Qu. 0.037 0.038 0.052 0.354 0.019 0.038 0.011 0.014 0.021 0.044 0.030 0.008 0.362 0.017 0.005 0.006 0.123 0.023 0.014 0.011 0.023 0.010 0.006 0.006 0.005 0.006
#> Max. 0.783 0.798 0.505 0.748 0.402 0.291 0.351 0.191 0.300 0.390 0.263 0.196 0.800 0.133 0.242 0.085 0.892 0.333 0.124 0.259 0.139 0.179 0.374 0.054 0.156 0.058The
SI.sig.bpobject is outputted from BayesPrism. We ran the BayesPrism with a single run by combining two technical replicates of Visium slides together (In the SpaceFold paper, we deconvolved them separately. As we observed that these two technical replicates were highly similar, we concatenated them to give better estimates in the 2nd Gibbs sampling of BayesPrism).Next we load the gene annotation file
features.tsv.gz(found from the Space Ranger output). This allow us to easily choose to either plot by gene ID or gene symbol. The file can be downloaded from https://figshare.com/ndownloader/files/36114461 .
gene.tab <- read.table("features.tsv.gz",sep="\t",check.names=F,header=F)[,1:2]
colnames(gene.tab) <- c("id",'symbol')
rownames(gene.tab) <- gene.tab[,"id"]
head(gene.tab)
#> id symbol
#> ENSMUSG00000051951 ENSMUSG00000051951 Xkr4
#> ENSMUSG00000089699 ENSMUSG00000089699 Gm1992
#> ENSMUSG00000102331 ENSMUSG00000102331 Gm19938
#> ENSMUSG00000102343 ENSMUSG00000102343 Gm37381
#> ENSMUSG00000025900 ENSMUSG00000025900 Rp1
#> ENSMUSG00000025902 ENSMUSG00000025902 Sox17- We start by constructing a SpaceFold object:
sf.obj <- new.sf(bp.obj = SI.sig.bp, feature= gene.tab)- One may also subset the sf.obj spots to subset on regions that are homogeneous on only one underlying dimension. For example, in our large intestine tissue, we found the upper tissues and lower tissues parts on the Visium slide seem to have slightly different cell type distribution, which may reflect varatiation of transcription along the proximal-distal axis of large intestine. To focus on the crypt axis only (rather than the distal-proximal axis), we subset the upper and lower part, and run SpaceFold separately on them. To do so, simply call the funtion:
#NOT run
sf.obj.part1 <- sf.obj[sf.obj$label=="part1"]
sf.obj.part2 <- sf.obj[sf.obj$label=="part2"]
# where label is a colname of sf.obj@meta to indicate the variable based on which to subset. - We can remove
SI.sig.bpfrom the R environment to save some memory.
rm(SI.sig.bp)
gc()
#> used (Mb) gc trigger (Mb) max used (Mb)
#> Ncells 7528739 402.1 12346808 659.4 9597161 512.6
#> Vcells 924879204 7056.3 2420693859 18468.5 2013357003 15360.7- Next we compute the background level for each cell type by using mixture models. We also define cutoffs for cell type fraction, theta, and the total reads of cell type over the selected genes (Znk). The final background will then be the maximum of the cutoff determined by the mixture model and the user-defined cutoffs.
sf.obj <- compute.background.level(sf.obj,
theta.cutoffs.user = 0.001,
Znk.cutoffs.user = 20,
posterior.cutoff = 0.7) #posterior value above which is assigned to the cluster. Default=0.7.
#> fitting mixture models on theta...
#> Current cell type: Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> fitting mixture models on Znk...
#> Current cell type: Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell- User can inspect number of spots selected for each cell type.
sf.obj
#> Data info:
#> Number of spatial spots: 4487
#> Number of genes: 3908
#> Fraction of cell types:
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> Min. 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.002 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
#> 1st Qu. 0.004 0.003 0.006 0.063 0.003 0.005 0.002 0.004 0.003 0.005 0.003 0.002 0.035 0.004 0.001 0.002 0.017 0.002 0.003 0.003 0.005 0.002 0.002 0.002 0.001 0.002
#> Median 0.011 0.010 0.016 0.196 0.007 0.011 0.005 0.007 0.007 0.015 0.008 0.004 0.094 0.008 0.002 0.004 0.039 0.006 0.007 0.006 0.011 0.005 0.003 0.003 0.002 0.003
#> Mean 0.049 0.052 0.048 0.227 0.021 0.031 0.013 0.012 0.020 0.039 0.022 0.007 0.199 0.014 0.006 0.005 0.153 0.018 0.011 0.009 0.017 0.008 0.005 0.005 0.005 0.005
#> 3rd Qu. 0.037 0.038 0.052 0.354 0.019 0.038 0.011 0.014 0.021 0.044 0.030 0.008 0.362 0.017 0.005 0.006 0.123 0.023 0.014 0.011 0.023 0.010 0.006 0.006 0.005 0.006
#> Max. 0.783 0.798 0.505 0.748 0.402 0.291 0.351 0.191 0.300 0.390 0.263 0.196 0.800 0.133 0.242 0.085 0.892 0.333 0.124 0.259 0.139 0.179 0.374 0.054 0.156 0.058
#> Total reads of each cell type :
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> Min. 0.240 2.108 2.724 8.752 1.512 2.088 0.108 0.948 1.824 2.968 0.572 1.072 3.296 1.844 0.596 1.360 5.172 1.352 0.776 1.412 2.040 1.384 1.204 1.148 0.756 1.104
#> 1st Qu. 13.758 6.234 12.234 280.864 5.822 17.496 5.388 7.664 5.728 23.670 4.334 5.874 75.304 12.250 2.880 4.310 71.176 10.900 12.628 7.456 14.236 6.342 5.078 3.844 2.184 3.152
#> Median 35.872 13.744 24.716 666.348 14.156 39.436 13.888 15.236 11.660 44.864 14.468 10.764 208.844 23.004 5.876 6.328 114.680 20.404 26.152 14.504 30.500 11.420 9.504 5.596 3.500 4.048
#> Mean 174.580 131.702 82.612 1292.645 42.058 144.589 56.914 37.981 35.478 90.892 129.679 29.684 1006.261 41.857 21.452 8.208 564.336 58.462 49.585 28.811 55.470 21.732 20.426 7.030 10.770 4.624
#> 3rd Qu. 111.976 45.844 61.982 1722.730 39.680 128.338 35.904 34.698 27.666 94.958 128.210 29.452 1405.994 46.726 12.456 9.672 329.224 58.750 56.058 30.856 65.746 22.296 18.836 8.270 6.958 5.356
#> Max. 11395.568 5206.788 2985.000 15032.564 987.392 3426.984 4003.808 1228.896 1303.548 1648.472 2133.264 1295.532 12672.224 787.912 2222.660 325.956 6929.324 2134.680 1760.776 669.100 1362.520 627.464 1949.016 123.648 1072.252 89.288
#> Number of spots above background for each cell type:
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> 813 858 651 1718 690 1180 425 524 807 809 1739 1009 1785 702 348 170 961 1228 483 250 777 579 224 83 334 11
#>
#> SpaceFold parameter.
#> $theta.cutoffs
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> 0.041834734 0.040299262 0.054212311 0.173113661 0.025987854 0.018407931 0.026216806 0.017803111 0.016356786 0.035378190 0.009513852 0.007373489 0.001000000 0.014737803 0.010907857 0.005373429 0.210660124 0.009750703 0.018435523 0.025208389 0.014597346 0.011833999 0.013053149 0.008206102 0.009461187 0.005404264
#>
#> $Znk.cutoffs
#> Paneth TA bottom zone enterocyte mid zone enterocyte Lgr5+ progenitor goblet cycling plasma goblet 1 Lgr5+ stem secretory progenitor goblet 2 macrophage top zone enterocyte tuft neurons/enteroendocrine cycling/GC B cell myofibroblast stromal 3 cDC/monocyte blood endothelium stromal 1/2 lymphatic T cell pDC glial resting B cell
#> 140.268 56.592 84.392 938.004 34.436 94.764 55.972 48.904 35.004 107.500 24.776 22.648 376.460 55.344 24.596 20.000 266.536 43.212 75.044 39.924 74.748 28.300 30.360 20.000 20.000 20.000
#>
#> $posterior.cutoff
#> [1] 0.7- Next, we perform data diffusion to learn the underlying crypt-villus
axis using
PHATE.
sf.obj <- run.phate(sf.obj,
if.invert=TRUE, # invert the final axis as the orientation is arbitruary
n.jobs=20) # run using 20 CPU threads - Let’s plot the top 5% of spots containing each cell type along the predicted crypt-villus axis:
# set the ordering of cell types (same as the paper)
myorder <- c("myofibroblast","glial","stromal 3","stromal 1/2","blood endothelium","lymphatic",
"Paneth","Lgr5+ stem","Lgr5+ progenitor","TA","bottom zone enterocyte","mid zone enterocyte","top zone enterocyte",
"secretory progenitor","goblet cycling","goblet 1","goblet 2","tuft","neurons/enteroendocrine",
"resting B cell","cycling/GC B cell","plasma","T cell","cDC/monocyte","macrophage","pDC")
plot.beeswarm(sf.obj,
cell.type.order= myorder,
pdf.prefix="output") #set pdf.prefix to your name. Default is output.
#> png
#> 2- Open file
"output_spacefold_beeswarm.pdf"in your working directory.
Gene expression cartography
- Cartography can be plotted for a group of cell types. To do so, we first define the cell type group of interest, and then merge their expression:
group.list <- list(enterocytes=c("bottom zone enterocyte","mid zone enterocyte","top zone enterocyte"))
sf.obj <- merge.cell.type (sf.obj= sf.obj,
grouping.list = group.list)Denoise cartography. This is a new function implemented by SpaceFold 2.0.
denoise.cartographyapplys data diffusion over the imputed SpaceFold axis to oversome sparsity of Visium data and further improve the resolution of cartography over the sliding window mean used by the v1.0 (the SpaceFold paper).Parameters of denoise.cartography
ka: the Kth nearest neighbor to which the distance is used to compute the sigma of Gaussian adaptive kernel. Smaller value favors preservation of local structure. Default=5.tansition: The diffusion time. Larger value yields greater smoothness. Default=1000.
sf.obj <- denoise.cartography (sf.obj= sf.obj,
ka=5,
tansition=1000)Visualize cartography
Task one: visualize bottom and top zone enterocyte marker genes from Moor et al. ( https://doi.org/10.1016/j.cell.2018.08.063 ). We plot 8 markers from each group.
Important parameters:
raw.or.norm: whether to plot raw count or raw count normalized by total expression of each cell type. The choice depends on the question of interest. To study the total expression, use “raw”. To study per cell expression, use “norm”. When plotting expression for a group of cell types we recommend “raw”, as the normalization factor of a group of cell types may be affected by the abundance of individual cell types within the group.denoise: whether to plot denoised cartography (SpaceFold v2.0) or sliding window mean (SpaceFold v1.0)show.raw: whether to show raw deconvovled expression (before smoothing). Default=FALSE. Can enable this if the axis is of low confidence for troubleshooting.overlay.hist: whether to show the histogram of cell type distribution. Default=TRUE.span: parameter for fitting the spline curve. Recommend to use 0.25 if denoise=TRUE, and 0.75 if denoise=FALSE.
top.zone.genes <- c("Apoa4","Apoc3","Clca4a","Krt20", "S100a10","S100a6","Slc25a22","Tmsb4x")
plot.cartography (sf.obj= sf.obj,
raw.or.norm="raw",
denoise=F,
selected.genes= top.zone.genes,
selected.cell.types="enterocytes",
show.raw =FALSE,
overlay.hist=TRUE,
span=0.75,
pdf.prefix="SI.topzone.v1")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levels
plot.cartography (sf.obj= sf.obj,
raw.or.norm="raw",
denoise=T,
selected.genes= top.zone.genes,
selected.cell.types="enterocytes",
show.raw =FALSE,
overlay.hist=TRUE,
span=0.25,
pdf.prefix="SI.topzone.v2")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levelsV1.0:
V2.0:
bottom.zone.genes <- c("Atp5a1","Atp5g3","Ccl25", "Cox4i1", "Cox6c", "Gpx2", "Lgals4" ,"Plac8")
plot.cartography (sf.obj= sf.obj,
raw.or.norm="raw",
denoise=F,
selected.genes= bottom.zone.genes,
selected.cell.types="enterocytes",
show.raw =FALSE,
overlay.hist=TRUE,
span=0.75,
pdf.prefix="SI.bottomzone.v1")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levels
plot.cartography (sf.obj= sf.obj,
raw.or.norm="raw",
denoise=T,
selected.genes= bottom.zone.genes,
selected.cell.types="enterocytes",
show.raw =FALSE,
overlay.hist=TRUE,
span=0.25,
pdf.prefix="SI.bottomzone.v2")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levelsV1.0:
V2.0:
Next, we plot the normalized expression of selected lymphatic marker Lyve1 and ligand genes in lymphatics and blood endothelium.
secreted_factors=c('Lyve1','Ntn1','Il33','Wnt2','Ccl21a','Rspo3','Reln')
plot.cartography (sf.obj= sf.obj,
raw.or.norm="norm",
denoise=T,
selected.genes= secreted_factors,
selected.cell.types=c('lymphatic','blood endothelium'),
show.raw =FALSE,
overlay.hist=TRUE,
span=0.25,
pdf.prefix="SI.secreted_factors.v2")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levels
plot.cartography (sf.obj= sf.obj,
raw.or.norm="norm",
denoise=F,
bin.by="equal.size",
selected.genes= secreted_factors,
selected.cell.types=c('lymphatic','blood endothelium'),
show.raw =FALSE,
overlay.hist=TRUE,
span=0.75,
pdf.prefix="SI.secreted_factors.v1")
#> Warning in brewer.pal(color.num, "Dark2"): minimal value for n is 3, returning requested palette with 3 different levelsV1.0:
V2.0:
Next, we plot the markers for Paneth (Lyz1) and stem cell (Lrg5) and Sox9 which expresses in Paneth, stem cell, progenitor and TA at different levels in a selected set of epithelial cell types.
plot.cartography (sf.obj= sf.obj,
raw.or.norm="norm",
denoise=F,
selected.genes= c("Lyz1","Lgr5","Sox9"),
selected.cell.types=c("Paneth","Lgr5+ stem","Lgr5+ progenitor","TA", "enterocytes"),
show.raw =FALSE,
overlay.hist=TRUE,
span=0.75,
pdf.prefix="SI.Lyz1_Lgr5_Sox9.v1")
plot.cartography (sf.obj= sf.obj,
raw.or.norm="norm",
denoise=T,
selected.genes= c("Lyz1","Lgr5","Sox9"),
selected.cell.types=c("Paneth","Lgr5+ stem","Lgr5+ progenitor","TA", "enterocytes"),
show.raw =FALSE,
overlay.hist=TRUE,
span=0.25,
pdf.prefix="SI.Lyz1_Lgr5_Sox9.v2")V1.0:
V2.0:
Noted that there are very few spots of Paneth and Lgr5+ stem mistakenly show up on the right portion of the axis. They represent spots containing a residual amount of those cell types. We can remove them by either redefine the background level by increasing the posterior cutoff, or simply define a hard cutoff on the SpaceFold axis. This is legitimate given known biology. We filter those spots using the
select.spotfunction as follows:oprepresents the relation with the existing selected set.
sf.obj <- select.spot (sf.obj= sf.obj, cell.type="Paneth", selected.idx= sf.obj@SpaceFold.axis<0.5, op="and")
sf.obj <- select.spot (sf.obj= sf.obj, cell.type="Lgr5+ stem", selected.idx= sf.obj@SpaceFold.axis<0.5, op="and")
sf.obj <- select.spot (sf.obj= sf.obj, cell.type="top zone enterocyte", selected.idx= sf.obj@SpaceFold.axis>0.5, op="and")- Each time spots are re-selected, we need to redo the cell type merging and denoising:
sf.obj <- merge.cell.type (sf.obj= sf.obj,
grouping.list = group.list)
sf.obj <- denoise.cartography (sf.obj= sf.obj, ka=5, tansition=1000)- Make the new plot:
plot.cartography (sf.obj= sf.obj,
raw.or.norm="norm",
denoise=F,
selected.genes= c("Lyz1","Lgr5","Sox9"),
selected.cell.types=c("Paneth","Lgr5+ stem","Lgr5+ progenitor","TA", "enterocytes"),
show.raw =FALSE,
overlay.hist=TRUE,
span=0.75,
pdf.prefix="SI.Lyz1_Lgr5_Sox9.v1")- V1.0:
